The two segments of the COS FUV detector are affected by fixed-pattern noise. This has two major causes: 1) small scale irregularities in the response of the detector and 2) shadows on the detector from the repeller grid wires. The latter are the stronger of the two, often causing decreases in the observed flux by as much as 20% over 50 pixels or more. Until now, we have flagged these regions and eliminated their contributions to the final, summed spectra. However, we have recently developed a flat field which corrects for the grid-wire shadows and begun to apply it to the processed data through the on-the-fly-recalibration (OTFR) as of April 6,2011. While pixels affected by grid wires are still flagged, their corrected values are included in the summed spectra. Figure 1 shows the effect of correcting the grid-wire shadows on a single G130M FUVB observation of the white dwarf, WD0320-539. This star has a relatively smooth continuum, making the corrections obvious. The upper (blue) spectrum contains grid-wire shadows (indicated by the vertical lines) which are corrected in the lower (green) spectrum. The affected regions are clearly much improved.
The reason we correct for the grid-wire shadows before addressing the remaining fixed-pattern effects is twofold. First, they are the largest single source of fixed-pattern noise. Second, because the shadows are perpendicular to the spectrum, their effect on the data is relatively insensitive to the location of the spectrum in the cross-dispersion direction. The remaining fixed-pattern noise is often strongly dependent on the spectrum location in the cross-dispersion direction and will require considerably more effort to fully characterize and correct.
With the grid-wire shadows corrected, the remaining fixed-pattern features create an effective noise which depends on both the detector segment (the two segments have different irregularities) and grating (G130M spectra are wider in the cross-dispersion direction and, therefore, smooth out small scale features). The table below gives our estimates of the maximum signal-to-noise (S/N) attainable from a single FP-POS exposure or by combining data from all four FP-POS settings, to increase it by a factor of 2. In either case, for the fixed-pattern errors to dominate the quadratic sum of fixed-pattern and Poisson errors, the exposure times must be large enough for the Poisson errors to be roughly half of the fixed-pattern errors. To attain higher S/N, it is necessary to use special analysis procedures, such as those discussed in the January 2011 Newsletter.
|Combining 4 FP-POS|
An ISR detailing the derivation of the flats, their application and their effects on the data is in preparation(Ely et al., 2011) and should be published within the next few months.